Optimasi Jumlah Cluster pada Algoritme K-Means untuk Evaluasi Kinerja Dosen

Abstract

Lecturers are one of the main actors in universities. Lecturer performance can affect the quality of a college. Because the quality of higher education is strongly influenced by the lecturers, the performance of the lecturers needs to be assessed. Lecturer performance can be assessed by evaluating the lecturer's performance in teaching. Lecturer performance can be evaluated by classifying student assessments of lecturers. Lecturers are assessed based on how the lecturer mastered the material, the lecturer's discipline in teaching, and the presentation of the material. The process of grouping lecturer scores can be done using the K-Means algorithm. K-Means is a popular clustering algorithm that performs well. K-Means requires a parameter that is K or the number of clusters. The importance of the number of clusters so that there is a need for optimization in determining the number of K. In this study, optimization was carried out using the Elbow method so as to produce the ideal number of groups of 4 groups. The results of the clustering evaluation calculated using the SSE were 54.4% which showed that the results were not optimal. The results of the cluster evaluation are not optimal due to the lack of data for the K-Means application.